“We trained an algorithm to decode pig sounds,” said Elodie Briefer, a co-scientist from the University of Copenhagen. against the Guardian† To do this, the team trained a so-called neural network, which attempts to simulate the properties of the brain.
any It happened Based on audio recordings and data of pigs in all positive and negative situations. In total, the scientists used 7,414 unique sounds from more than 400 animals to train the algorithm.
Get to know the feelings
Various sounds are associated with the animals’ feelings, which range from very positive, such as interactions with family members and suckling by the mother, to very negative, such as fighting and castration. In the end, the algorithm was able to correctly guess whether the sentiment was negative or positive 92 percent of the time.
App for farmers
The next step is to turn the research into a practical application. “Now we need someone to develop the algorithm into an application that farmers can use to improve the welfare of their animals,” Bender said.
The researchers hope that this technology will lead to a platform that will allow farmers to better monitor the psychological state of animals.
